IMPORTANT: If your linkage works on large number of records with large number
of columns or if you are enablind results desuplication (in either available 
form), you may need to increase the memory available for FRIL. Go to FAQ 
section of the webpage for instructions.


Version 2.1.5 (current version)
Fixed bug when using FRIL in Batch mode (Null pointer exception).


Version 2.1.4
Fix: Fixed bug in selection of location for results file

Version 2.1.3
Fix: Fixed problem in fril.bat (spaces in JAVA_HOME caused FRIL not to start)
Update: Notification if compared columns are not saved into output file

Version 2.1.2
Update: FRIL saves match score to file containing records being duplicates
        (if deduplication is enabled)

Version 2.1.1
New feature: Dynamic analysis tables show tooltips with full string value
Fix: Close window button works in dynamic analysis windows
Update: By default, all attributes used in FRIL will be saved into output file
Update: FRIL attempts to search for file with data in a few locations
Update: Greatly reduced memory consumption when results deduplication is used
Update: General memory-oriented optimization

The only difference between versions 2.1 and 2.1.1 is that the latter
uses a new version of openscv library used to read CSV files.

Version 2.0:
1. New feature: Information in status bar about linkage results
2. New feature: User can turn on linkage summary (click on icon in status bar)
3. New feature: When reading text files (CSV, plain text), file encoding is recognized
                FRIL supports plain ASCII, utf-8, utf-16 and utf-32 encodings.
4. Fix: Street address distance restores all parameters properly
5. Fix: Replace string converter fully supports empty string
6. Fix: Strata join did not save the strata info and had bug in strata join results dedupe
7. Update: EM method 'suggests' percentage of data source for sampling
8. Update: Empty value slider (linkage/dedupe) triggers analysis window update
9. Update: Automated support of JDBC drivers. New drivers should be placede in
                jdbc directory (see FAQ on FRIL webpage for details).
10.Results viewer shows row numbers

Version 1.0 patch7:
1. New feature: Street address distance
2. New feature: Manual decision module can be closed, rejecting or accepting all linkages
3. New feature: Sort enabled in table in dynamic analysis window
4. Fix: Memory leak in results deduplication
5. Fix: Memory leak in sorted neighborhood search method
6. Fix: Possible exception when removing item from condition (dedupe condition)
7. Update: Updated deduplication condition configuration window

Version 1.0 patch6:
1.  New feature: Deduplication supports weightsb and acceptance level
2.  New feature: Support for comparing empty values
3.  New feature: Empty values can be set for each of attributes used in FRIL (Data source config.)
4.  New feature: Prefix of attribute value can be used as a blocking variable
5.  New feature: Attribute names are validated against correctness (Data source config.)
6.  New feature: Manual review of linkages (Linkage configuration)
7.  Fix: Restarting blocking search method saved some values from previous round
8.  Fix: Summary statistics failing after multiple runs of linkages using blocking search
9.  Fix: Deduplication process more responsive to cancel button
10. Fix: Bugfixes in autosave/backup of configuration
11. Fix: SNM could miss reporting one record to minus files in certain circumstances
12. Fix: Deduplication in result savers could omit saving some not joined records for large linkage result
13. Update: New blocking configuration panel (deduplication and linkage modes)
14. Update: Miscellaneous changes in GUI

Version 1.0 patch5:
1. New feature: Save removed duplicates from data source into files
2. New feature: Linkage/Deduplication mode
3. New feature: Memory usage monitoring
4. New feature: Save data from data view
5. Viewer for minus files
6. Warning when viewing the reuslts for the fisrt time
7. Improved summary information after linkage/deduplication
8. Improved "Show fields summary" window

Version 1.0 patch4:
1. Fixed bug in results deduplication (error when no linkage results were produced)
2. Fixed labels in data source wizard
3. Added "File->New" option
4. Added filtering of data source
5. Fixed: minus reporting for blocking search method
6. Fixed: results deduplication could identify false duplicates (bug in recID of DataRow).
7. Added results analysis (results analysis dialog)
8. Added option of manual review of results duplicates (a dialog with questions appears)
9. Fixed: If necessary, results deduplication adds records to minus files
10. Fixed: Few other minor bugs fixed

Version 1.0 patch3:
1. Added Jaro-Winkler distance support
2. Added deduplication of linkage results (options in results savers config)
3. Added autorecovery of unsaved configuration if application crashes
4. Added option to open configuration of other datasource (datasource configuration)
5. New feature in Tools menu - deduplicate file
6. Improved Performance/memory management
7. Improved GUI organization

Version 1.0 patch2:
1. Fixed bug in Q-Grams distance (calculation of score could give wrong results for some cases)
2. Fixed bug in SNM (reporting of not joined records could omit some data)
3. Fixed bug in removing many list entries at once using "minus" button
4. Improved linkage optimization and memory management
5. Improved support for dynamic analysis
6. Improved multiple sheets support in Excel data source

Version 1.0 patch1:
1. New feature: script editor in converters
2. Bugfixes: Nested Loops, Sorted Neighborhood Method
3. Bugfixes: User interface (join/data source configuration)
4. Bugfixes: Excel data source
5. Bugfixes: Date Distance - after and before ranges were switched
   WARNING: If you used this distance you might need to revise your configuration

Version 1.0
1. Improved cancel responsivity (for load configuration)
2. New Feature: Support for multicore architectures
3. Improved progress bars (real progress is shown)
4. Improved memory management
5. New Feature: Excel files support
6. New Feature: Window for configuration JDBC connection
7. New Feature: Configuration window for properties
8. New Feature: Linkage histogram
9. Corrected SNM method (e.g. when date distance is chosen, data is sorted based on dates comparison)
10. Minor bugfixes
11. New Feature: Data source deduplication

Version: 0.6
1. Feature of copy join between strata
2. Info on strata name in saved results
3. 'Add all columns' feature in configuration of data source
5. 'Rename column' feature in configuration of data source
6. Various bugs in user interface fixed

Version: 0.5
1. Split converter now supports more than one split character
2. Feature of 'Edit converter' in datasource configuration
3. New converter: Replace converter
4. Fixed not-joined summary (saving in the same directory as result, saving only column names)
5. Revised numeric and date distance functions
6. Feature of 'Edit result saver'
7. New feature: multi-level converters
8. New feature: Data stratification
9. Moved to Java 1.6 (better memory management)
10. Order selection in SNM method now moved to last step of wizard
11. Updated feature: New progress window
12. Updated feature: Configuration loading and join preparation are separated now
13. Minor bugfixes across the application

Version: 0.3
1. Limit on number of iterations in EM
2. New features in numeric distance
3. Added support for distance between dates/time
4. Improved sampling in "Show examples" when adding condition
5. Added information on errors when dynamic analysis is used
6. Fixed bug: dynamic analysis window did not close when switching between different distance functions
7. Corrected minor bugs in user interface